It's this massive thing, right? Alabama sent two football players over to do TikTok Rush with the sorority we're working with. They're two best players they sent over there. The colleges are leaning in because they've got a gap to fill, right? They've got a $30 million deficit. They got a hit now that they've never had before. They're leaning in. I work, I'm meeting with the AD of Alabama in two weeks to go through this. We're also going to announce, keep out for the end of October. Concerts work really well with music. We're not Live Nation, so we don't need to make a 20% margin. We can break even. Imagine, say, Morgan Wallen at Bryant-Denny Stadium in April would be pretty cool, wouldn't it?
By the way, if you want tickets to that as we go forward and grow this out, I don't care if you have a Black Card, a Platinum Card, a Sapphire Card, get in the back of the line. How much Yay Alabama product? How much Yay Ohio State? How much Yay Oregon product did you buy? You buy more, you go front of the line and a discount on the ticket. Imagine being able to sit on the floor and watch Morgan Wallen for $100, but because you bought $300 in Yay Alabama gear. We break even on that. Free marketing. Just an idea. It might not be him. Maybe it is, but that's what we're doing. We have our loyalty programs. Think about Sephora and Ulta. Our loyalty is music. The only way you get access to tickets and better prices, you support us.
The sororities, we've already talked to them. We said, here's what you got to spend as a sorority on our product. If you spend that much, first in line. Sorority girls, Alabama has 18 or 19 sororities, 800 girls each. Right? Imagine 60 universities across the country, Greek system, student athletes, families, alumni, and all these influencers coming on on top of it. Pretty powerful. It's a good margin stack. I know you got to go. I got to go.
So I'm, Mike Zemetra. I'm the CFO of Veritone. I've been with Veritone since October of 2020. Veritone at a macro level, we are in the AI space. Nice little buzzwords. We've been around since 2014. The company was founded by two serial entrepreneur brothers. On their last voyage of selling a company, they sold to Google. They came up with the concept of Veritone and what it is today. Today we're about a little over 400 employees. We've got offices around the world. We are in the U.S. We're in predominantly all of Europe, with offices in the U.K., Germany, and France, Australia, and then we have some satellite offices in India and in Israel. We've got about 3,000 customers today, and we went public in 2017. What does Veritone do? At a simplistic level, we take unstructured data.
That's everything that you consume from your cell phone, video, audio, emails, and we make it intelligent. If you look at the rate of unstructured data today, it's everything that we do from cell phones to Zoom calls, et cetera. We are heavily in two verticals. One is the commercial sector, and I'll go into these in more detail. We have a very growing public sector, which is really state and local law enforcement, as well as federal government with the Department of Defense and the Department of Justice. In terms of unstructured data, this is just to give you a snapshot of the gravity of how much we process. Last year we processed over 58 million hours of media. If you think about it in the context, that's every day processing the Walt Disney Company's film library. That's the power of our platform. It's over 10 PB.
That's from here to the moon several times. I'll explain in more detail. We do utilize unique AI models, whether they're machine learning or large language models. We have over 850 in our platform. As a snapshot, again, unstructured data is video, audio, text, images, email, things that you can't structure today. We've got a platform we call aiWARE. aiWARE, at its simplistic form, has over 26 levels of cognition built into it. What's cognition? That's everything a human does. That's object detection, so identifying it, identifying a human. We have both biometric and non-biometric ways of identifying people. We have translation. We can take in real time any language, whether it's coming from speech to text, translate it speech to text. It's very powerful. The beauty of this platform, and I'll explain it a little bit, it's a little bit like Microsoft Office.
We've got the platform on top, and then we've built applications like Excel, Microsoft Word, et cetera, underneath, which we're in market today, and I'll go into those in more detail. As I mentioned before, this API, it sits on top of all cloud providers, AWS, Microsoft Azure. It can API any feed in terms of data, so it's very malleable. It can both sit on the cloud and on site. Some of our, because of intel for intelligence reasons, we have to be deployed in private environments like GovCloud or for the Department of Defense and their own tenant. It can go both in tenant and can sit on the cloud. As I mentioned before, a little bit like Microsoft. This is just a little bit. We've got over 3,000 customers.
There's a lot to put on this list, but we're on the commercial side, it's everyone from the NCAA to ESPN. I'll explain in terms of what we do for them in more detail. On the public sector, it's everyone from the Department of Defense. We just announced a huge contract with the Air Force in June of this year, to state and local law enforcements, which is everyone from LAPD all the way down to local state and sheriff departments. In terms of the commercial sector, as I mentioned before, we're both in commercial and in public sector. The commercial sector has been, you know, our mainstay, and the public sector is growing. I'll give you a use case here in terms of applications. We help ESPN program SportsCenter. If you watch the end of every SportsCenter, you're going to see our logo appear. How do we do that?
We're literally ingesting across all their properties in real time feeds. Whether it's, you know, the Herd or whether it's SportsCenter, what they're talking about, and that just allows them to program. We also deal with our advertisers, so their advertisers understand how to monetize. We also help them in terms of archiving. They have access, we're able to index their entire 30+ year library. If they want to pull up every Tiger Woods putt where he's fist pumping and the Nike logo in the background, they can literally do it on their keyboards. They used to have about 250 interns in Connecticut that did this for them. They no longer have that. We're an efficiency product for them as well. We also do content management. The NCAA, everyone's familiar with the NCAA, of course. We are with the exclusive rights over all their digital content.
So whether it's the March Madness or it's the college football playoffs, anything in terms of monetization comes through our platform because our platform is so massive in terms of what it can digest. We literally host the likes of CBS News, Bloomberg, the NCAA, et cetera, which allows third parties to come in, find the content, and then monetize it. A good example of that is the movie Air with Michael Jordan. They had NCAA footage in it. They had to license that through us in our platform. Then we get to the government side of the business. On the government side, we basically go the whole spectrum of public safety. From identifying bad people, we can do that biometrically, non-biometrically. Think about the Capitol riots, and you had CCTV footage, you had about 1,000 cell phones, you had cameras inside of the Capitol building.
We can actually take all that data, ingest it in our system, time correlate it, so we can follow the person in the Nike hat with the black shoes, where they started and where they ended up. Why is that important? A lot of states have laws that prevent you from using pure identification because it's not perfect. It could criminalize somebody who's particularly innocent. Actually having this, you know, not only can it catch the person in the Nike shirt, but we can actually identify the dent on the door if a criminal decided to paint their car. That's how powerful this system is. We call it iDEMS. Recently we announced a deal with the Air Force, where we are the backend on their investigative capabilities, which is basically the beachhead to go across the entire Department of Defense. There are 17 divisions of it.
We are super excited about where this product's going to go. Then we get to one of our most recent growth initiatives. In the back half of last year, we launched a product, again, an application across aiWARE, called Veritone Data Refinery. We refer to it as VDR. VDR at its simplistic ingest content, but we're not talking about a small amount of content. We're talking about hundreds of thousands of hours of content. Digitize it, index it, and allows hyperscalers, this is the AWSs, the Metas, the OpenAIs of the world, to feed that content and train their large language models. Why is that important? If you actually look at, go to the next slide, where training data is becoming essential. All these models have scaled through what I call static data, text, or that's the World Wide Web, or images. Now they're moving on to video.
You want to train a robot to run like an NFL athlete, you need video. You want to train bespoke images just on subtleties of light and shadow, you need video. You don't just need a thousand hours of video, you need hundreds of thousands of hours of video to train something on a bespoke level. There's just not many companies capable of being the aggregator of that content, being able to digitize it and feed it into these systems. When we went out and launched this in the back half of last year, we had sub $1 million of bookings. Today our bookings and our pipeline's over $40 million. We see this as a big market. Today it's a little north of $3 billion. We see it going to $17 billion by 2032. We hope to be a big part of this ecosystem.
In terms of our last quarter, we did about $24 million in revenue. It was basically flat year over year, but that's a little bit misleading. Our core software, when you exclude our Hire business, which I'll explain in a second, grew over 45% year- over- year. This quarter we're projecting to do growth of 30% on aggregated revenue. Our bottom line is expected at our midpoint to improve over 40%. We are still in the process of growth and burning capital, but that year over year improvement is really trajecting us to get to an operating cash flow break even by the back half of next year. We're super excited about it. As I mentioned before, we have over 3,000 customers. We're all over the world. In terms of ARR, which is our core software, we've got a little over $62 million of ARR, so sticky revenue.
Our customer retention is world class. I mean, we look at it on a dollar basis, and it's well in the high 90th percentile. It says we have cash and cash equivalents on there. That's a little bit misleading because we just raised over $100 million of equity in the past two months. We can skip that. I mean, this just shows optically our quarterly revenue. We are now going into the growth phase, and this is going to be even more pronounced when we get to the third quarter, which we're going to be announcing here in early November. I want to save questions. Yes.
I saw you did a, it looks like a shelf offering maybe for $75 million recently.
Yep, last week.
You just mentioned over $100 million in.
We did another offering in September as well, a little over $25 million.
Okay.
Yep.
How are you planning on reporting that down?
We've got about $125 million of debt. It's a little over $30 million of term debt, and we've got about $91 million in change of convertible debt. Just expect us to use a good portion to improve our liquidity position. Yeah. Yep.
Me.
Yes.
Can you characterize your Air Force contract a little bit?
Yeah, so.
Do you have a list of allies? You know, is it one of a kind or?
Yeah, if you go back to Doge and the Trump administration, they're about centralization. Centralization just isn't centralizing the whole thing. It's centralizing the Department of Defense. It's also centralizing their software stack. They have multiple, if whether you're the.
Are you going to run rates about 100?
Yeah, we're projected to do between $108 million and $115 million this year. Last year we did about $92 million.
I guess you're not profitable at that level, but I'm trying to get a line around how's it turning into a profitable business.
Yeah, our gross margins are north of 60% today, and our costs are actually down year- over- year. There's not a lot of incremental costs we need, you know, and I mentioned before in the third quarter, at the midpoint, we're going to achieve over 30% revenue. It's really the growth with extreme cost management. Yep, yep.
What's the blending of the margins, software?
Sorry?
What's the gross margins on the software?
We don't break it out between software, but on a blended basis, it's north of 60%. I think we were north of 65% last quarter, yeah. Yep.
We've seen concern about legal challenges for privacy concerns. It's slightly scary when you see the problems occurring.
Yeah, the beauty is we are experts in copyright and infringement just for the customers. That's why a lot of the hyperscalers are coming to us to make sure they do this clean. Yeah. Yep.
Outside research?
Yeah, today we're covered by Ladenburg, UBS, God, you stumped me on this one. D. Boral and H.C. Wainwright, and just a little birdie told me that Needham's going to be initiating coverage this week. You bet. Yep.
You have a software business, which last quarter grew 40% year- over- year.
Our core software, but it's over 45%, yeah.
Over 45%. What's the workflow business?
We take our software and we have a managed service business, which is predominantly that content licensing business, which is very consumption-based. It's not, it uses our software, but it is a consumption-based model. That grew about 6% last quarter. Then we've got a small bit of a representation business that declined mostly from macro, but that's really what we'll call non-core from software.
What percentages are called non-core right now?
Yeah, I think we did about $17.5 million in software, and the residual, so $7 million, $8 million, is on the managed service side.
When I saw your quarterly revenues, it seems like there's a big seasonality, with the cost per quarter being the highest. Would that continue?
There's not much seasonality in our business. If there is seasonality, it's on the content licensing side because we do sponsor live events. So Augusta National, everyone knows the Masters. They're actually utilizing our software on site to do real-time content management. We've got a room full of 100 people just like this. They're utilizing our software to take real-time feeds from a hundred different cameras, and then it's published on XYZ.com. Similarly, we do that for the USTA, so at the U.S. Open for tennis. That creates a little bit of seasonality, but, you know, and then in March we have March Madness, and then in December, in the fourth quarter, we have football. It actually ends up all kind of streaming itself out. Yeah. Yep.
Is your AI technology, is it all in-house or do you license?
aiWARE, the platform, is completely homegrown built. We do use over 800 different types of machine learning. The beauty of that is we're not in a race to compete with OpenAI or Gemini. We can use best in class and just interchange them. We do have third parties in the platform, yeah.
Sorry, can you give me density? Is the AI density in regards to engagement, and therefore is it also enough?
Sorry?
Is the AI density in regards to engagement, and therefore is it also enough?
Yeah, I mean, our velocity is definitely increasing in terms of the number of. Is that your question?
Are you under pressure to go with the console system with the AI?
I mean, no, not necessarily. Like I mentioned, the majority of our revenue is on cloud. Of course, AWS went down this week, so who knows, but we can scale pretty efficiently in that environment. Yes.
Do you foresee our software growing at these rates?
Yeah, wait for Q3. It's going to be bigger, yeah.
Yeah, it sounds like, from what you're saying, it sounds like we could, but in the near future after that?
I mentioned that the TAM here is enormous. AI is enormous. Our SAM is enormous. Unfortunately, our lawyers don't like the way that we put our SAM on a slide, so they won't let us put it into a deck, but it's billions and billions of dollars. Yeah. Yep.
I'm going to vary myself.
Sure.
Are there any comps within Palantir that you can read about that might be a little too?
That's a good question. I mean, I think Palantir is a little bit of the opposite of us. They're pretty heavily involved in the public sector, more so than they are in the commercial, albeit growing, whereas we're more heavily involved in the commercial sector and not so much in the public sector, but growing. I don't, I wouldn't say, I mean, we have competed against them head to head. Obviously, their velocity and their size is much larger than ours, by a multiple. We go head to head with Axon, which is, you know, the thousand-pound gorilla. It just kind of depends on the use case, and we'll mostly see them in the federal space, not necessarily on the commercial side. Yeah. Yep. All right. Oh, one more. Yep.
Can you add to, I mean, add to the iPad?
Yeah, our founders came from an advertising background, mostly in the radio side. We had a full-fledged media advertising agency, predominantly on the audio side, but we sold that business last year. It was non-core to us. There's a little residual amount of that, but it's very small. Yeah.
Do you foresee being able to get to like the end of next year without raising further capital?
Sorry, without raising capital?
Yes.
I think it'll depend on the velocity of growth and, you know, being opportunistic. I mean, our stock is up pretty substantially since the beginning of the year, so we're always going to be opportunistic. Yeah, yep.
If you were to raise capital, would it be a follow-on offering or a shelf offering, or how would you do that?
Stay tuned. Yeah, there's a lot of efficient ways to raise capital. We did have an ATM that was up previously, but we'll talk more about that in a couple of weeks during Veritone's call.
Are you inquisitive at all?
We were very inquisitive. We were doing an acquisition a quarter. We really stopped that a couple of years ago to focus on the core. I mean, we've taken out north of $40 million for us, which is over 40% of our burn today, of cost out of the company to really get us to be more lean focused where we are today.
You're focused really on that debt that you're getting?
We've publicly said that part of the proceeds would be used to pay down debt, yeah.